Dendritic cells (DC) play an important role in the induction of T-cell responses. We hypothesize that the hampered antiviral T-cell response in chronic hepatitis B patients is a result of impaired dendritic cell function. In this study, we compared the number, phenotype and functionality of two important blood precursor DC, myeloid DC (mDC) and plasmacytoid DC (pDC), of chronic hepatitis B patients with healthy volunteers. No differences in percentages of mDC and pDC in peripheral blood mononuclear cells were observed between chronic hepatitis B patients and healthy controls. The allostimulatory capacity of isolated and in vitro matured mDC, but not of pDC, was significantly decreased in patients compared to controls. Accordingly, a decreased percentage of mDC expressing CD80 and CD86 was observed after maturation, compared to controls. In addition, mDC of patients showed a reduced capacity to produce tumor necrosis factor α after a stimulus with synthetic double-stranded RNA and interferon γ. Purified pDC from patients produced less interferon α, an important antiviral cytokine, in response to stimulation with Staphylococcus aureus Cowan strain I than pDC isolated from controls. In conclusion, mDC and pDC are functionally impaired in patients with chronic hepatitis B. This might be an important way by which hepatitis B virus evades an adequate immune response, leading to viral persistence and disease chronicity. (HEPATOLOGY 2004;40:738–746.)
Hepatitis B virus (HBV) infection represents an enormous health problem worldwide. More than 350 million people are chronically infected with HBV and are at risk to develop liver cirrhosis or hepatocellular carcinoma. Until now it is unclear why an individual develops a chronic carrier state. However, an inadequate immune response of the host is thought to play a critical role in the chronicity of the infection.1 To recover from an acute HBV infection, both a strong humoral and a strong cellular immune response are required. During an acute infection, patients exhibit a multispecific and polyclonal cytotoxic T-cell (CTL) response and a strong type-1 T helper cell response.2, 3 Such HBV-specific T-cell responses are generally undetectable in chronic patients.4, 5
Dendritic cells (DC) represent the most potent antigen-presenting cells and thus play an important role in the induction of specific T-cell responses.6 Functional defects in DC could therefore be an important mechanism of the virus to evade host immune responses. In several chronic viral infections, such as human immune deficiency virus 1 and hepatitis C, impaired function of DC has been demonstrated.7–11 For HBV, evidence of an impaired DC function in chronic patients exists as well.12, 13 Data on the functionality of DC in chronic HBV patients were obtained with in vitro-generated monocyte-derived DC (moDC). Recently, Osugi et al. showed several differences between moDC and the myeloid DC present in vivo.14 Therefore, it would be more accurate to isolate and study the functionality of DC precursors present in vivo. Two major DC precursors are the myeloid dendritic cells (mDC) and plasmacytoid dendritic cells (pDC).15, 16 Myeloid DC are characterized by the absence of surface expression of lineage markers and the presence of markers such as CD1c (blood dendritic cells antigen [BDCA] 1) and CD11c. In response to bacterial compounds or CD40 ligand, mDC can produce large amounts of interleukin (IL) 12, and they require the presence of granulocyte-macrophage colony-stimulating factor (GM-CSF) for survival.15
Plasmacytoid DC, on the other hand, are characterized by the negativity of lineage markers and the presence of BDCA2, BDCA4, and CD123 (IL-3 receptor α-chain), require the presence of IL-3 for survival, and can produce high amounts of type I interferon (IFN), IFN-α and IFN-β, on exposure to viruses and bacterial components such as CpG oligonucleotides.17–19 The release of IFN-αβ initiates a cascade of events that eventually lead to the elimination of the virus. It can act on several cells and turn on biochemical pathways that restrict viral replication and render host cells resistant to further viral infection.20 Moreover, type I IFN is frequently used as therapy for chronic HBV patients and leads to disease remission in approximately 20% to 35% of cases.21, 22 Therefore, the role of type I IFN-producing pDC could be of great interest in the pathogenesis of chronic HBV infection.
The aim of this study was to determine the number, phenotype, and functionality of mDC and pDC precursor subsets in the peripheral blood of chronic HBV patients and to compare these characteristics with those of healthy volunteers.
Peripheral heparinized blood samples were obtained from 30 patients with chronic hepatitis B (Table 1). All patients were negative for antibodies against human immune deficiency virus, hepatitis C, and hepatitis D. None of the patients were treated for chronic HBV infection or received any other medication 6 months previous to blood sampling. All patients had biopsy-proven chronic hepatitis with minimal to moderate fibrosis. The median serum HBV DNA load was 1.1 × 108 geq/mL (range, 1.0 × 103–1.2 × 1010) and median alanine aminotransferase level (ALT) was 48 U/L (range, 9-243). Fifteen patients were HBV envelope antigen (HBeAg)-positive, and 15 patients were HBeAg-negative and had antibodies to HBeAg. A control group, matched for age, sex, and race, comprised 19 healthy subjects who had no evidence of exposure to HBV (HBV surface antigen [HBsAg]-negative). The study was approved by the local ethics committee, and all patients and controls in the study gave informed consent before blood donation.
Table 1. Patient Characteristics
All Patients (n = 30)
Controls (n = 19)
Abbreviations: n.t., not tested; n.a., not applicable.
Serum HBsAg, HBeAg, and anti-HBe were determined quantitatively using the Abbott IMX system (Abbott Laboratories, North Chicago, IL) according to the manufacturers instructions. Serum HBV DNA was determined using an HBV monitor assay (Roche Applied Science, Penzberg, Germany; detection limit, 1 × 103 geq/mL). When the serum HBV DNA was below 1 × 103 geq/mL HBV DNA, the assay was repeated using an in-house developed TaqMan polymerase chain reaction (detection limit, 373 geq/mL) (Applied Biosystems, Foster City, CA).23 TaqMan polymerase chain reaction was also used for determination of HBV DNA in the DC subtypes. DNA was extracted from a pellet of 2 × 104 DC. The pellet was resuspended in 190 μL double distilled water, to which 10 μL of a known amount of internal control, consisting of seal herpes virus, was added. This material was extracted on a MagnaPure LC isolation system (Roche Applied Science) using the total nucleic acid isolation kit. Twenty μL of the isolated nucleic acid was used to detect HBV DNA. Values for the internal control had to be within a range previously determined within the laboratory.
Analysis of mDC and pDC Subsets in Peripheral Blood.
Peripheral blood mononuclear cells (PBMC) were obtained by Ficoll-Isopaque gradient centrifugation. PBMC were incubated for 5 minutes with human immunoglobulins (Octagam, 1.3 mg/mL; Octapharma, Lachen, Switzerland) in phosphate-buffered saline containing 1% wt/vol bovine serum albumin before adding specific antibodies to determine the percentage of mDC and pDC. For mDC, PBMC were incubated with a cocktail of FITC-conjugated monoclonal antibodies to the lineage markers CD3, CD14, CD16, CD19, CD20, and CD56 (Becton Dickinson, San Jose, CA), anti-BDCA1 (CD1c)-PE (Miltenyi Biotec, Bergisch Gladbach, Germany), and anti-CD11c-APC (Becton Dickinson). To determine the percentage of pDC, PBMC were incubated with the FITC-conjugated lineage cocktail, anti-BDCA4-PE and anti-CD123-biotin, followed by a secondary step with streptavidin-PerCP (Becton Dickinson). As controls, cells were stained with corresponding isotype-matched control monoclonal antibodies. Stained cells were analyzed using a 4-color flow cytometer (FACScalibur, Becton Dickinson) and CellQuest software (Becton Dickinson).
Isolation of mDC and pDC.
Myeloid DC (BDCA1+) and pDC (BDCA4+) were isolated from PBMC by positive immunomagnetic selection using the mini-MACS system (Miltenyi Biotec) according to the manufacturer's instructions. Briefly, BDCA1+ mDC were isolated by incubating CD19-depleted cells with a PE-conjugated monoclonal antibody to BDCA1, followed by anti-PE magnetic beads and separation over a MS-column. In addition, BDCA4+ pDC were isolated after incubation with a PE-conjugated monoclonal antibody to BDCA4, followed by anti-PE magnetic beads. Cells were resuspended in culture medium consisting of RPMI 1640 (Bio Whittaker, Verviers, Belgium) containing 10% fetal calf serum (Hyclone, Logan, UT). The isolated mDC and pDC were analyzed for purity by flow cytometry as mentioned in the previous paragraph (BDCA1+ and lineage markers− CD11c+ for mDC; BDCA4+ and lineage markers− CD123high for pDC) and only used if more than 90% pure.
Expression of Cell Surface Molecules on the Surface of mDC and pDC by Flow Cytometry.
Freshly isolated mDC and pDC (1 × 104 cells) were incubated with CD80-FITC (Immunotech, Marseilles, France), human leukocyte antigen (HLA)-DR-PerCP (Becton Dickinson), and CD86-APC (Becton Dickinson). As a control, cells stained with corresponding isotype-matched control monoclonal antibodies were used. Cells were analyzed on the flow cytometer and the mean fluorescence intensity and percent positive stained cells was determined.
Analysis of T-cell Stimulatory Capacity of mDC and pDC in a Mixed Lymphocyte Reaction.
Purified mDC and pDC were matured for 24 hours in 96-well flat bottom culture plates at different concentrations (1.25, 2.5, 5, and 10 × 103 cells/200 μL) in culture medium containing IL-1β (50 ng/mL; Strathmann Biotech, Hannover, Germany) and tumor necrosis factor (TNF) α (25 ng/mL; Strathmann Biotech). As a growth supplement, GM-CSF (500 U/mL; Leucomax, Novartis Pharma, Arnhem, The Netherlands) was added for mDC and IL-3 (10 ng/mL; Strathmann Biotech) was added for pDC. The next day, culture supernatant was removed and nylon-wool–purified T cells from a normal healthy volunteer (1.5 × 105 cells/200 μL) were added to the DC. After 5 days, cell proliferation was assessed by the incorporation of [3H]thymidine (Radiochemical Centre, Amersham, Little Chalfont, UK), 0.5 μCi/well was added, and cultures were harvested 18 hours later. Phytohemagglutinin 5 μg/mL (Murex, Paris, France) was added to T cells as a positive control.
Cytokine Production of Stimulated mDC and pDC.
Purified mDC were stimulated at a concentration of 4 × 104 cells/200 μL in 96 flat bottom Costar plates (Costar, Cambridge, MA) in culture medium with synthetic double-stranded RNA, polyriboinosinic-polyribocytidylic acid (poly[I:C], 20 μg/mL; Sigma-Aldrich, St. Louis, MO), in combination with recombinant human IFN-γ (1000 U/mL; Strathmann Biotech) and in the presence of GM-CSF (500 U/mL). Purified pDC were stimulated at a concentration of 2 × 104 cells/200 μL with Staphylococcus aureus Cowan strain I (SAC; 75 μg/mL; Calbiochem, San Diego, CA) in the presence of IL-3 (10 ng/mL). Cells were cultured at 37°C and 5% CO2 in a humified incubator. Supernatants were harvested after 24 hours. The level of IL-12p70 (Diaclone, Besançon, France) and IFN-α (Biosource International, Nivelles, Belgium) were determined by standard enzyme-linked immunosorbent assay (ELISA) according to the manufacturer's instructions. The levels of TNF-α, IL-6, and IL-10 were determined by specific solid-phase sandwich ELISA, using pairs of monoclonal antibodies and recombinant cytokine standards from Biosource International, as previously described.24, 25
Data are expressed as mean ± SEM, unless indicated otherwise. Data were analyzed with SPSS 11.5 for Windows (SPSS, Chicago, IL) using the Mann-Whitney test to compare variables between 2 independent groups. In all analyses, a P value of less than .05 was considered statistically significant. Correlations were determined using the Spearman correlation test. HBV patients were compared to healthy controls. To analyze the results on the basis of serum viral load and ALT, we categorized all HBV patients according to the following: HBV DNA greater than 108 geq/mL (high; n = 16); HBV DNA 105 to 108 geq/mL (intermediate; n = 5); and HBV DNA less than 105 geq/mL (low; n = 9); and ALT normal (upper limit of normal [ULN] = 35 U/L; n = 11); ALT normal to 2 × ULN (intermediate; n = 9); and ALT greater than 2 × ULN (high; n = 10).
Myeloid and Plasmacytoid DC Percentages in Peripheral Blood Are Similar in Chronic HBV Patients and Healthy Controls.
To determine whether a chronic HBV infection affects the frequencies of mDC and pDC in peripheral blood, the percentages of both cell types in PBMC were determined by flow cytometry. Myeloid DC were recognized in peripheral blood by staining with a cocktail of lineage markers (CD3, CD14, CD16, CD19, CD20, and CD56) and anti-BDCA1 (Fig. 1A). Plasmacytoid DC were characterized by staining with anti-BDCA4 and anti-CD123 (Fig. 1B). No significant differences in the frequencies of mDC and pDC were observed between chronic HBV patients and healthy controls (Fig. 1C).
Myeloid DC of Chronic HBV Patients Are Inhibited in Their Capacity to Express Costimulatory Molecules on in vitro Maturation.
Freshly isolated DC precursor populations from peripheral blood of patients and controls express only low amounts of the costimulatory molecules CD80 and CD86. After stimulation for 24 hours in the presence of IL-1β and TNF-α, the percentages of mDC and pDC expressing CD80 and CD86 and the mean fluorescence intensities of the positive cells were significantly increased. Maturation of mDC was significantly more increased compared to pDC, as shown in the flow cytometry histograms of 1 representative healthy control (P < .001; Fig. 2). No differences were observed in the mean fluorescence intensities of CD80 and CD86 on mDC from patients and controls. However, the percentages of mature mDC expressing CD80 and CD86 after stimulation were significantly reduced in the patient group (P < .05; Table 2). The percentages of pDC expressing CD80 or CD86 of patients and controls were similar both before and after stimulation. A significant increase in mean fluorescence intensity of CD86 on pDC was found in the HBV patient group after stimulation (P < .05). The HLA-DR mean fluorescence intensities were significantly increased on mDC and pDC after stimulation (P < .001; Fig. 2 and Table 2). No difference in HLA-DR mean fluorescence intensity was found when comparing patients to controls.
Table 2. Phenotypic Characteristics of mDC and pDC
The Allostimulatory Capacity of mDC of Chronic HBV Patients Is Impaired.
To assess whether the T cell-stimulating capacity of the two dendritic cell subsets is affected in chronic HBV patients, we studied the ability of DC to stimulate T cells in the allogeneic mixed lymphocyte reaction. Mature mDC and pDC (obtained as described in the previous paragraph) were added at different concentrations to T cells of a healthy third party. Fig. 3A shows that mDC of patients were less efficient in inducing T-cell proliferation than mDC isolated from controls at all ratios tested (P < .05). Plasmacytoid DC were capable of inducing T-cell proliferation but were less potent than mDC. This was also reflected by their lower expression of HLA-DR and costimulatory molecules, CD80 and CD86. No difference was found in the T-cell stimulatory function of pDC from patients and controls (Fig. 3B). Background proliferation of T cells only was less than 500 cpm. The proliferation of phytohemagglutinin–stimulated T cells was similar in both groups (data not shown).
Impaired TNF-α Production by mDC and IFN-α Production by pDC of Chronic HBV Patients.
Myeloid DC are primary producers of IL-12 and pDC are primary producers of IFN-α. Chronicity of HBV infection is thought to be caused mainly by a reduced type-1 T helper cell response, possibly due to reduced IL-12 and/or IFN-α production by DC. To investigate the capacity of the 2 subtypes of DC to produce cytokines, mDC were stimulated with poly(I:C) and IFN-γ; to induce high IL-12p70 production, and pDC were stimulated with SAC for high IFN-α production. Twenty-four hours later, culture supernatants were harvested and analyzed for the production of several cytokines using specific ELISAs. The production of IL-12p70, IL-10, and IL-6 by mDC was not significantly different when comparing patients with controls (Figs. 4A-C). However, the secretion of TNF-α in the supernatant of stimulated mDC isolated from the patients was significantly reduced compared with healthy controls (P < .05; Fig. 4D). IFN-α was not detectable in the culture supernatant of stimulated mDC (data not shown).
SAC-stimulated pDC of patients showed a significantly impaired production of IFN-α compared to controls (P < .05; Fig. 5C). IL-10 production by pDC was increased in the patients, although not significantly (P = .1; Fig. 5B). There was no significant difference in IL-6 and TNF-α production of SAC-stimulated pDC between controls and patients (Figs. 5A and D). Only low amounts of IL-12p70 were detectable in SAC-stimulated pDC (data not shown).
Dendritic Cell Function in Relation to Viral Load and ALT Levels.
Analysis of the obtained results of the HBV patients in relation to viral load and ALT showed that the capacity of mDC to produce TNF-α was significantly reduced in patients with low viral load compared to patients with high viral load (P < .05; Fig. 6A). No relation was observed between TNF-α production and ALT levels.
The IFN-α production of pDC was significantly decreased in patients with high ALT compared to patients with normal ALT (P < .05; Fig. 6B). Although all patients with high ALT exhibited HBV DNA levels above 108 geq/mL, no relation was found between serum viral load and IFN-α production.
Expression of costimulatory molecules before and after maturation, allostimulatory capacity, and capacity to produce cytokines by mDC and pDC were not related to viral load and ALT.
To exclude the possibility that the observed effects were the result of liver inflammation rather than a specific effect of the virus, we studied a separate group of patients with chronic inflammatory liver disease of nonviral origin—primary biliary cirrhosis (n = 2) and hemochromatosis (n = 4)—and an ALT comparable to that of chronic HBV patients (median 40 U/L; range, 24-110 U/L). The mDC isolated from these patients showed similar expression of costimulatory molecules and similar allostimulatory capacity to those of healthy controls. Furthermore, the production of IFN-α by pDC was not reduced in these patients compared to healthy controls (data not shown).
HBV DNA Was Detectable in Dendritic Cell Precursor Subpopulations of Chronic HBV Patients.
A functional impairment of the dendritic cell subsets could be caused by the presence of virus in the DC. To investigate this, HBV DNA was determined in the DC by TaqMan polymerase chain reaction. HBV DNA was detected in mDC of 15 of 30 chronic HBV patients. In 12 patients, HBV DNA was present in pDC. The level of HBV DNA was significantly higher in pDC than in mDC: 1.2 × 104 ± 2.9 × 103 geq/mL versus 6.7 × 103 ± 2.2 × 103 geq/mL, respectively (mean ± SEM; P < .01). The presence of HBV DNA in mDC and pDC was correlated with serum HBV DNA values: (Spearman correlations r = 0.64, P < .05 and r = 0.60, P < .05, respectively). HBV DNA was not detected in any of the control DC tested. No correlation was observed between HBV DNA presence in mDC and suppressed allostimulatory capacity and decreased CD80 and CD86 expression of matured mDC (data not shown).
DC are professional antigen-presenting cells that are extremely potent in initiating a primary immune response. The weak or absent T-cell responses found in chronic HBV patients could be the result of a defect in the dendritic cell compartment. Numbers and/or functionality of dendritic cell subsets in the blood may be affected by the presence of the virus. The present study showed that percentages of mDC and pDC in peripheral blood of chronic HBV patients and controls were similar. On the other hand, the functionality of the 2 dendritic cell precursor subsets was affected in HBV-infected patients compared to healthy controls.
In vitro matured mDC isolated from chronic HBV patients exhibit an impaired allostimulatory capacity compared with mDC from healthy controls. It is unlikely that this difference is caused by the difference in the HLA class II allele mismatches, since the allostimulatory capacity of pDC between patients and healthy controls was similar. The reduced T-cell stimulatory capacity may be due to the decreased up-regulation of the costimulatory molecules CD80 and CD86 after in vitro maturation of mDC from HBV patients. However, pDC of patients showed an increased expression of CD86 after in vitro maturation, while there was no difference in T-cell stimulatory capacity of pDC of patients and controls. This suggests that other factors, such as costimulatory or inhibitory molecules, play a role in the decreased allostimulatory capacity. For example, CD40 has been shown to serve as a costimulus for T-cell activation, and also CD40 ligation is a critical step in the final maturation of DC into fully competent antigen-presenting cells.26
Besides the expression of costimulatory molecules or inhibitory molecules, the capacity of mDC to produce certain cytokines is also important for the T-cell stimulatory capacity. Myeloid DC of HBV patients were impaired in their ability to produce TNF-α. The capacity of mDC to produce TNF-α was decreased significantly in patients with low viral load. Sheron et al. showed that serum TNF-α levels and the in vitro production of TNF-α production by lipopolysaccharide-stimulated PBMC was lower in HBeAg-negative patients with low viral load.27 TNF-α has been shown to contribute to allostimulation; it can act as an autocrine growth factor for dendritic cell-induced T-cell proliferation.28 Furthermore, TNF-α plays an essential role in the maturation of DC.29 We demonstrated that addition of exogenous TNF-α induced maturation of DC. This was manifested by the increased expression of costimulatory molecules after maturation with IL-β and TNF-α. However, the maturation of mDC was reduced in HBV patients compared to healthy controls. The capacity of mDC to produce TNF-α was most prominent in patients with low viral load, while the allostimulatory capacity of mDC was comparable for all HBV patients. This indicates that there may be an additional factor yet to be determined that also plays a role in the reduced allostimulatory capacity of the mDC of HBV patients.
In contrast to previous studies using moDC, we showed that IL-12 production by mDC was not different between HBV patients and controls. In previous studies, a reduced allostimulation corresponded with a reduced IL-12 production in chronic HBV patients.12, 13 In a study by Lohr et al., the reduced T-cell response induced by moDC from chronic HBV patients could be restored by exogenous IL-12.30 Taken together, the results for IL-12 production show that mDC are clearly different from moDC, as also previously shown by Osugi et al.14
In addition to the impaired function of mDC in chronic HBV patients, we found a strongly reduced IFN-α production by pDC from chronic HBV patients. Furthermore, a trend toward up-regulation of the cytokines IL-6, TNF-α, and IL-10 was observed in these patients, indicating an imbalance in cytokine production by the pDC of HBV patients. IFN-α is a pleiotropic modulator of host resistance and plays a critical role in the induction of an antiviral state.19, 31 In other viral infections, such as human immune deficiency virus 1, a reduced capacity of pDC to produce IFN-α also has been shown.32, 33 Furthermore, it has been proposed in human immune deficiency virus infection that pDC control viral replication through production of IFN-α.34 This antiviral effect of IFN-α may also play a role in chronic HBV infection, since we found a more reduced IFN-α production in patients with active disease (high ALT and high viral load). Recently, IFN-α has also been shown to play a role in plasma cell differentiation and subsequent antibody production.35 Therefore, reduced IFN-α production may influence the production of HBV-specific antibodies. Exogenous IFN-α therapy may restore the disturbed cytokine environment and could be successful, especially in those patients with impaired endogenous IFN-α production. This issue is subject of further study.
Our results indicate that both circulating dendritic cell subsets in peripheral blood of patients with chronic HBV infection are affected in their function. This functional impairment of DC was not observed in patients with chronic inflammatory liver disease of nonviral origin. Therefore, our findings indeed appear related to HBV and are not solely the consequence of inflammation of the liver. The dendritic cell dysfunction in chronic HBV patients may be caused directly by viral infection of DC, interfering with signaling pathways involved in maturation and cytokine production. We showed the presence of HBV DNA in both dendritic cell subsets. Others have shown that PMBC and purified cell subsets contain HBV DNA and also viral intermediates, suggesting viral replication.36, 37 Moreover, viral particles and HBV replication intermediates were found in moDC of chronic HBV patients,12, 13 and direct infection of DC has been described for other viruses.38–40 HBV may also interact with DC through binding to the cell surface receptors, as previously shown for human immunodeficiency virus and hepatitis C infection.41 From our results, we cannot exclude the possibility that the presence of HBV DNA in both dendritic cell subsets reflects attachment of the virus to the cell surface or uptake of the virus as a biological function of DC. The virus may also indirectly affect the dendritic cell subsets, for instance by infection of other cells that in turn produce cytokines, causing a change in cytokine environment in the liver or lymphoid organs. Further research should determine the mechanism by which HBV interferes with dendritic cell functions and whether the HBV DNA we found in the DC represents replicating virus.
The question remains whether chronic HBV infection in patients is the cause of a preexisting dysfunction of the dendritic cell subtypes or whether the infection itself causes the dysfunction of DC. Recently, a study on hepatitis C in chimpanzees suggested that the dendritic cell impairment is rather a consequence of persistent and active hepatitis C virus infection associated with disease progression.42 Research studying the functionality of DC before and after successful therapy for HBV may provide answers for this issue.
In conclusion, our results show that both mDC and pDC are functionally impaired in chronic HBV patients. This could be one of the reasons for the absent or weak T-cell responses and subsequent continued HBV replication in these patients.
The authors wish to thank S. Pas, Department of Virology, Erasmus MC, for the determination of HBV-DNA levels in serum and DC and Dr. L. J. W. van der Laan for comments and advice.